1. About
  2. Data
  3. Insights
  1. About
  2. Data
  3. Insights
  1. Subscriber Access

Case Study

Interactions and Interests: Collaboration Outcomes, Competitive Concerns, and the Limits to Triadic Closure

By Antonia Hunter

"My study breaks new ground in our limited understanding of how venture capital syndication relationships can shape LP investment decisions. I would not have been able to do this research without Preqin, who were very helpful in addressing my needs with custom data extracts on LP investments."

Pavel Zhelyazkov
Assistant Professor
Hong Kong University of Science and Technology

Pavel Zhelyazkov, Assistant Professor at HKUST, presents a summary of the full paper using Preqin LP investment data

My study explores how the syndication relationships between venture capital (VC) firms affect the pattern of investments by limited partners (LPs). LPs – large institutional investors such as pension funds, endowment plans, foundations, and financial institutions – play a critical role in the entrepreneurial finance ecosystem by providing capital to the VC firms that support start-up companies; however, to date, there has been little research about how LPs select the VC firms that they invest in.

In particular, I focus on the mediating role of VC firms that have received prior investments from an LP but also have had syndication relationships with other VC firms. Such mediating VC firms can play an important role in introducing their LPs to their syndication partners, and by assisting their LPs in their due diligence process by answering questions about difficult-to-observe aspects of their syndication partners’ capability and behaviors. At the same time, they have mixed motives: while they may want to help their syndication partners in accessing financing from their LPs, they may also be concerned about connecting their LP to a potential competitor that can supplant them in the future. Thus, the question arises: under what conditions would an LP be more or less likely to invest in syndication partners of VC firms in which the LP has previously invested?

To answer this question, I merged Preqin’s database of LP investments into VC firms with Thomson Reuters’ database of VC syndication relationships. Drawing on the investment activity of 610 different LPs making 4,668 initial investments in VC firms over the 1997-2007 period, I examined what distinguished the LP-VC investment pairs that did happen from others that could have happened but did not. I controlled statistically for many factors that can explain matching – such as the VC’s overall performance, its geographical distance from the LP, as well as its differences from other VCs that the LP had previously invested in – and zeroed in on the effects of VC-mediated indirect ties between the focal LP and focal VC.

Preqin is very flexible and willing to work with researchers. The team were able to assist with custom extracts from the Preqin database, providing me with one comprehensive dataset that was extracted to my specification. Data provided was clean and consistent, and the ability to conduct meetings with Preqin research staff in which they provided early feedback to my emerging findings was also really appreciated.

I found that while on average LPs were more likely to invest in VCs to which they were indirectly connected via an intermediary VC, the effect of such VC-mediated ties was dependent on two major factors. First, what was the outcome of the syndication relationships between the mediating and the focal VC? If they were successful, then the likelihood of the mediating VC’s LPs investing in the focal VC was dramatically enhanced; however, failures between the two VCs would make the mediating VC’s LPs less likely to invest in the focal VC than in similar VCs without any connections. In other words, if a VC is disappointed in its dealings with a syndication partner, then negative information would spill over to its own LPs, and that syndication partner is unlikely to get future investments from them, even if its overall investment performance is held constant.

Second, even a successful VC-mediated tie would not increase the likelihood of matching between an LP and a focal VC if the mediating VC has competitive concerns: in other words, if the mediating VC is afraid that the focal VC would replace it as the favored investee of the LP. Such competitive concerns are especially acute under two conditions: 1) if the focal VC has higher attractiveness than the mediating VC (i.e. because of superior performance); and 2) if the mediating VC is easily replaceable by the focal VC (i.e. the two VCs have very similar investment profiles, so they are easily substitutable with minimal implications for the LP’s portfolio). Indeed, I find that when both these conditions hold, even successful VC-mediated ties reduce the likelihood of LP investments in the focal VC, presumably because the mediating VC would be working actively against such connections. Furthermore, the competitive concerns of mediating VCs are well founded: in cases where one of its LPs establishes a tie with one of its more attractive syndication partners, a mediating VC is less likely to receive a re-investment from that same LP.

In summary, my study breaks new ground in our limited understanding of how VC syndication relationships can shape LP investment decisions. It also has clear managerial implications for fundraising VC firms. First, in prioritizing LPs to approach, VC firms are well advised to focus on LPs that are likely to hear good news about them during due diligence (i.e. because they have ties to VCs with which the focal VC has had successful experiences). Second, a focal VC cannot assume that even syndication partners with which it has had a successful relationship would help it access their LPs if those syndication partners are concerned about losing their investors to a more attractive rival.

The full paper is published in the Administrative Science Quarterly: Zhelyazkov, P. I. (2018). "Interactions and interests: Collaboration outcomes, competitive concerns, and the limits to triadic closure." Administrative Science Quarterly 63(1): 210-247.

Available online here.